Learning an Implicit Physics Model for Image-based Fluid Simulation

2025-08-11

AI Summary

This work addresses the challenge of synthesizing physically plausible 4D (3D + time) fluid animations from a single static fluid image—a task hindered by neural networks’ limited capacity for intuitive physical modeling. We propose a physics-guided neural framework that incorporates implicit physical constraints via Navier–Stokes equation-based loss terms, enabling joint kinematic and dynamic optimization. To decouple appearance and dynamics, we introduce a feature-level disentangled 3D Gaussian representation. Our method integrates monocular depth estimation, differentiable rendering, and end-to-end physics-informed neural network (PINN) training. The resulting animations achieve high visual fidelity while significantly improving physical consistency, temporal stability, and multi-view controllability. Quantitative and qualitative evaluations demonstrate that our approach outperforms state-of-the-art methods in both physical plausibility and perceptual realism.

Technology Category

• Computer Vision: Low Level & Physics-based Vision
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Humans possess an exceptional ability to imagine 4D scenes, encompassing both motion and 3D geometry, from a single still image. This ability is rooted in our accumulated observations of similar scenes and an intuitive understanding of physics. In this paper, we aim to replicate this capacity in neural networks, specifically focusing on natural fluid imagery. Existing methods for this task typically employ simplistic 2D motion estimators to animate the image, leading to motion predictions that often defy physical principles, resulting in unrealistic animations. Our approach introduces a novel method for generating 4D scenes with physics-consistent animation from a single image. We propose the use of a physics-informed neural network that predicts motion for each surface point, guided by a loss term derived from fundamental physical principles, including the Navier-Stokes equations. To capture appearance, we predict feature-based 3D Gaussians from the input image and its estimated depth, which are then animated using the predicted motions and rendered from any desired camera perspective. Experimental results highlight the effectiveness of our method in producing physically plausible animations, showcasing significant performance improvements over existing methods. Our project page is https://physfluid.github.io/ .

Problem

• Predict physics-consistent 4D fluid motion from single images
• Overcome unrealistic animations from simplistic 2D motion estimators
• Enforce physical principles via neural networks for fluid simulation

Innovation

• Physics-informed neural network predicts motion
• Feature-based 3D Gaussians capture appearance
• Navier-Stokes equations guide physical consistency